Abstract:Network-on-Chip (NoC) is a scalable on-chip communication architecture for the NN accelerator, but with the increase in the number of nodes, the communication delay becomes higher. Applications such as machine learning have a certain resilience to noisy/erroneous transmitted data. Therefore, approximate communication becomes a promising solution to improving performance by reducing traffic loads under the constraint of the acceptable maximum accuracy loss of neural networks. It is a key issue to balance the result quality and the communication delay for approximate NoC systems. The traditional approximate NoC only considers the node-to-node approximation-based dynamic traffic regulation. However, the dynamically changing traffic patterns across different nodes, different times, and different applications lead to a huge search space, which makes it hard to explore an optimal global approximation solution. In this paper, we propose a quality model for different neural networks, which presents the relationship between the quality loss and the data approximate rate. Then, a hierarchical approximate scheme optimized with reinforcement learning (HAS-RL) is proposed and we reduce the complexity of the HAS-RL by reducing the state space and action space, which will reduce the resource overhead as well. After that, we embed a global approximate controller in the NoC system, in which we deploy a policy network trained with the offline reinforcement learning algorithm to adjust the data approximate rates of each node at run time. Compared with the state-of-the-art method, the proposed scheme reduces the average network delay by $13.5\%$ while their accuracies are similar. The proposed HAS-RL only causes an additional area overhead of $1.24\%$ and power consumption of $0.77\%$ compared with the traditional router design.

Work in Progress: ACAC: An Adaptive Congestion-aware Approximate Communication Mechanism for Network-on-Chip Systems

Work in Progress: ACAC: an Adaptive Congestion-aware Approximate Communication Mechanism for Network-on-Chip Systems

Improving the MAC layer performance of Ad Hoc Networks by congestion control algorithms

On Performance Optimization and Quality Control for Approximate-Communication-Enabled Networks-on-Chip

Congestion Control for High Dynamic Ad Hoc Networks: Price Cooperation and Receding Optimization

Congestion Avoidance, Detection and Alleviation in Wireless Sensor Networks

Adaptive Congestion Control for Application Specific Networks-On-Chip

Massive Access For Machine-Type Communications In Backhaul-Constrained Heterogeneous Networks

Run-time Accelerate Channel for Communication-Aware Network-on-Chip

Adaptive Harvest-then-cooperate: Delay-aware Wireless Powered Communication Networks.

Latency Criticality Aware On-Chip Communication

Machine Learning-based Low Overhead Congestion Control Algorithm for Industrial NoCs

CACC: A Congestion-Aware Control Mechanism to Reduce INT Overhead and PFC Pause Delay

PACC: Perception Aware Congestion Control for Real-time Communication

Forward Active Networks Congestion Control Algorithm and Its Performance Analysis

A Hardware/Software Co-Design Methodology for Adaptive Approximate Computing in Clustering and ANN Learning

PACC: A Proactive CNP Generation Scheme for Datacenter Networks

Transaction-Aware Network-on-Chip Resource Reservation

Dynamic and Traffic-Aware Medium Access Control Mechanisms for Wireless NoC Architectures

HAS-RL: A Hierarchical Approximate Scheme Optimized with Reinforcement Learning for NoC-Based NN Accelerators

Re-architecting Congestion Management in Lossless Ethernet.